fix: use docker compose down (not stop) to properly tear down staging

stop leaves the container/network in place, blocking port rebind.
down removes everything cleanly.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

.github/workflows/deploy.yml

@@ -260,12 +260,9 @@ jobs:
 
       - name: Stop old containers
         run: |
+          docker compose --profile staging-go down 2>/dev/null || true
           # Kill legacy container names from pre-rename era
-          docker rm -f meshcore-staging meshcore-staging-go 2>/dev/null || true
-          # Kill current container names
-          docker compose --profile staging-go stop staging-go 2>/dev/null || true
-          docker rm -f corescope-staging-go corescope-staging 2>/dev/null || true
-          fuser -k 82/tcp 2>/dev/null || true
+          docker rm -f meshcore-staging meshcore-staging-go corescope-staging 2>/dev/null || true
           sleep 2
 
       - name: Start staging on port 82